Return to search

Hybrid Multi-Objective Optimization of Left Ventricular Assist Device Outflow Graft Anastomosis Orientation to Minimize Stroke Rate

A Left Ventricular Assist Device (LVAD) is a mechanical pump that is utilized as a bridge to transplantation for patients with a Heart Failure (HF) condition. More recently, LVADs have been also used as destination therapy and have provided an increase in the quality of life for patients with HF. However, despite improvements in VAD design and anticoagulation treatment, there remains a significant problem with VAD therapy, namely drive line infection and thromboembolic events leading to stroke. This thesis focuses on a surgical maneuver to address the second of these issues, guided by previous steady flow hemodynamic studies that have shown the potential of tailoring the VAD outflow graft (VAD-OG) implantation in providing up to 50% reduction in embolization rates. In the current study, multi-scale pulsatile hemodynamics of the VAD bed is modeled and integrated in a fully automated multi-objective shape optimization scheme in which the VAD-OG anastomosis along the Ascending Aorta (AA) is optimized to minimize the objective function which include thromboembolic events to the cerebral vessels and wall shear stress (WSS). The model is driven by a time dependent pressure and flow boundary conditions located at the boundaries of the 3D domain through a 50 degree of freedom 0D lumped parameter model (LPM). The model includes a time dependent multi-scale Computational Fluid Dynamics (CFD) analysis of a patient specific geometry. Blood rheology is modeled as using the non-Newtonian Carreua-Yasuda model, while the hemodynamics are that of a laminar and constant density fluid. The pulsatile hemodynamics are resolved using the commercial CFD solver StarCCM+ while a Lagrangian particle tracking scheme is used to track constant density particles modeling thromobi released from the cannula to determine embolization rated of thrombi. The results show that cannula anastomosis orientation plays a large role when minimizing the objective function for patient derived aortic bed geometry used in this study. The scheme determined the optimal location of the cannula is located at 5.5 cm from the aortic root, cannula angle at 90 degrees and coronal angle at 8 degrees along the AA with a peak surface average WSS of 55.97 dy/cm2 and stroke percentile of 12.51%. A Pareto front was generated showing the range of 9.7% to 44.08% for stroke and WSS of 55.97 to 81.47 dy/cm2 ranged over 22 implantation configurations for the specific case studied. These results will further assist in the treatment planning for clinicians when implementing a LVAD.

Identiferoai:union.ndltd.org:ucf.edu/oai:stars.library.ucf.edu:etd-7698
Date01 January 2019
CreatorsLozinski, Blake
PublisherSTARS
Source SetsUniversity of Central Florida
LanguageEnglish
Detected LanguageEnglish
Typetext
Formatapplication/pdf
SourceElectronic Theses and Dissertations

Page generated in 0.0011 seconds